Re: tighten input to float4/float8/oid

Tom Lane wrote:
> I think this is wrong.  We silently accept leading whitespace in
> (IIRC) all the numeric datatypes, and I believe we should accept
> trailing whitespace too.

(Sorry, I had misremembered your suggestion -- you had earlier said
that the spec probably allows for leading and trailing whitespace.)

I think there's a case to be made that we shouldn't accept either
leading or trailing whitespace, but it seems too late to go down that
path: it isn't worth breaking backward compatibility over, for one thing.

So if we're going to continue to accept leading whitespace, it seems
only reasonable to consistently accept trailing whitespace as well.

> Minor stylistic gripe here: errdetail and errhint messages are
> supposed to be complete sentences

Thanks for catching that, I'll fix it before applying.

I'll apply the patch this evening without the trailing whitespace
change; if there's a consensus that allowing trailing whitespace is
the best course, I'll post a separate patch for that shortly.

-Neil